"FoodPorn is your passport to a world where words morph into pictures of irresistible, exaggerated food indulgence" -ChatGPTThis Lora was trained to produce pictures of the most over-the-top delicious foods. It is intended for images in which the food is the focus, and will tend to overwrite non-food topics in your prompt. It should work a little bit for food generally, but will respond best to foods it was trained on. Here are some statistics about the labels that appeared in its training set, along with their frequencies:foodporn": 124, "savory": 92, "meat": 62, "cheese": 50, "dessert": 33, "pasta": 33, "fruit": 19, "chocolate": 19, "mexican": 17, "asian": 18, "cream": 18, "vegetables": 18, "american": 16, "potato": 15, "pastry": 12, "bread": 12, "hamburger": 11, "egg": 11, "shrimp": 10, "fried": 9, "dumpling": 8, "taco": 7, "waffle": 6, "wrap": 6, "italian": 5, "quesadilla": 3, "rice": 3.So the easiest way to prompt this lora is to just make a comma-separated list of the items you want. For example, "foodporn, savory, hamburger, pasta, potato, fried &lt;lora:FoodPorn_v2_:1&gt;" could be a prompt.While this Lora was trained on the SD 1.5 base model, I strongly recommend using it with the model <a target="_blank" rel="ugc" href="https://civarchive.com/models/18208?modelVersionId=68551">rMadArt</a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/18208?modelVersionId=68551">,</a> which was used for most of the images here. It doesn't work well with non-photorealistic models. I also recommend the negative embedding <a target="_blank" rel="ugc" href="https://civarchive.com/models/87781?modelVersionId=94126">boring_e621_v4</a>, which prettifies images further and the <a target="_blank" rel="ugc" href="https://civarchive.com/models/58390/detail-tweaker-lora-lora">Detail Tweaker</a> lora, which can adjust the level of detail shown in the images.

Files with SHA256: dfc8b3bdb064ec99c06e1a0e3aa0cbab08c00d6bb3310337aa3dd65d40b9aaa6